EWSMLFS: Explainable Weighted Stacked Ensemble Machine Learning with Recursive Feature Selection for Higgs Boson Particle Detection and Classification

Abstract

This research paper addresses the critical task of detecting and classifying Higgs Boson Particles (HBP) in high-energy physics research. We propose an explainable weighted stacked ensemble Machine Learning (ML) approach with Recursive Feature Selection (RFS) to achieve accurate results while ensuring model interpretability. By leveraging ensemble learning, we combine multiple Base Models (BM) in a stacked ensemble framework, assigning weights based on their performances to enhance prediction accuracy. We use RFS to identify the most relevant features to improve interpretability, reducing dimensionality and helping with a clearer understanding of the underlying physical processes. Our experiments on the HBP measurements dataset show our approach outperforms baseline models while maintaining transparency. We evaluate our model using key metrics, including accuracy, precision, recall, and F1 score. We analyze the interpretability of the ensemble and identify the most important features contributing to the classification process. Our results indicate that the proposed approach strikes an optimal balance between accuracy and interpretability. The combination of weighted stacked ensembles and RFS provides valuable insights into the HBP detection process. Scientists can gain a deeper understanding of the underlying physical phenomena, enhancing the reliability and trustworthiness of the particle classification results.